Ontologically grounding appearances in experience: Transcendental Idealism according to Anja Jauernig's The World According to Kant

In The World According to Kant Anja Jauernig defends a forthrightly idealist reading of Kant on which things in themselves exist independently of human minds while appearances (objects in space and time) are fully mind-dependent intentional objects of human experience. In these remarks I will concentrate more on our disagreements than on our agreements, as we philosophers typically do in these contexts. But let me say at the outset that I agree with a great deal of Jauernig's conclusions (having argued for similar ones myself) and I am pleased to see the resolutely idealist reading of Kant get such a full-throated and unapologetic defense. One of the merits of Jauernig's approach is that, by focusing on Kant's metaphysics and setting his epistemology to the side (deferred to a future book), she is able to address directly issues in the foundations of Kant's metaphysics that often go unexplored (e.g. whether appearances supervene on transcendental reality.) Without further ado, zu den Sachen selbst! TI1-Strong. (a) Empirical objects are not things in themselves in the transcendental sense, that is, they are not mind-independent; (b) rather empirical objects are Kantian outer appearances, that is, they are essentially intentional objects of outer experience as characterized in Kant's theory of experience, and thus, fully mind-dependent. (p. 357)2 Full mind-dependence. An entity E is fully mind-dependent if, and only if, all of E's determinations and other ontological ingredients are strongly mind-dependent. Strongly mind-dependent being. The being of an entity E is strongly mind-dependent if, and only if, necessarily, if E has being, there is a finite mind M, and E has being in virtue of the fact that M has [or could have] a possibly complex representation whose presentational content includes E.3 Putting these together, an entity E is a Kantian outer appearance if and only if, necessarily, if E has being, there is a finite human mind M and E has being in virtue of the fact that M has or could have a possibly complex representation, namely experience, whose presentational content includes E. An analogous principle holds for inner appearances (objects of inner experience, states of empirical minds), but I will focus on outer appearances (appearances in space).4 In Jauernig's terminology, the experience whose content grounds appearance is said to ‘ontologically specify’ appearances.5 My primary question is, what is the relevant notion of experience such that experience, so understood, ontologically specifies appearances? The case I want to focus on is what I will call ‘empirical properties’: properties that are not determined by the transcendental or metaphysical principles of nature alone.6 Appearances must have empirical properties, for they do not only obey the transcendental principles of the CPR and the metaphysical principles of the MFNS. They also have determinate sizes, shapes, and motions, masses, etc., and they obey particular natural laws that are not wholly grounded in transcendental and metaphysical principles. These sizes shapes, motions, and laws are governed by the transcendental and metaphysical principles but are not fully grounded in them alone.7 To put this point epistemically, if I have complete knowledge of the transcendental and metaphysical principles, I still cannot infer a priori their determinate sizes, shapes, motions, masses, laws, etc. Jauernig first turns to Kant's invocation at A110 (and elsewhere) of the ‘one and the same universal experience’ which different so-called ‘experiences’ are said to be parts. But what is this ‘one experience’ and how does it ontologically specify appearances? Jauernig characterizes it as follows: ‘a science-assisted empirically and transcendentally grounded account of the empirical world that is shared, or at least shareable, by all properly functioning members of the community of human minds’ (p. 82). This seems right to me, as does Jauernig's next move: she considers and then rejects as a candidate for the experience that ontologically specifies appearances ‘the final shareable science-assisted empirically and transcendentally grounded account of the empirical world that stands at the end of all scientific and observational progress’ (p. 82). Call this Final Experience. Final Experience cannot ontologically specify appearances because Final Experience is impossible. No experience can represent the unconditioned totality of empirical conditions of empirical objects, and since every empirical object has conditions that are themselves further empirical objects, no experience will represent every empirical object.8 Final Experience is at most a regulative ideal to which we asymptotically approach.9 An ew-account at time t is a shared, or at least sharable, account of the empirical world that includes a list of all general laws and all specific empirical laws of nature that have been specified by science so far; a list of general commonsensical descriptions of causal laws proposed so far that are inductively based on observations and govern various different kinds of phenomena; a record of all appearances that have been perceived to date as well as of all appearances that have been inferred so far on the basis of known laws and the available empirical evidence; and a collection of empirical judgments and judgments of experience made so far concerning the determinations of these appearances, including their relations to each other. (p. 362; cf. p. 95)10 She then considers whether what ‘ontologically specifies’ appearances is a possible ew-account or an actual one. Call these views Possible and Actual, respectively. She rejects Possible, the ontological specification of appearances by possible ew-accounts. Although I disagree with her specific argument against Possible, I agree with her that it must be rejected.11 Consequently, I will not discuss it further here. However, Jauernig overlooks a third possibility: ontological appearances are specified counterfactually by the ew-accounts we would have if we were to experience objects under certain conditions (e.g. if we were to institute a regress of experiences guided by the regulative ideal of Final Experience).12 Jauernig never explores this kind of view, and thus never explains her reasons for rejecting it. Instead, having rejected Possible (correctly, I think), she assumes the only alternative is Actual. If we define actual experience […] as the conjunction of all ew-accounts that are actual […], the same point can be put by saying that appearances exist and have all of their determinations and other ontological ingredients […] in virtue of being represented in actual experience. (p. 102–3; cf. p. 362)13 Because ew-accounts are something like scientific or proto-scientific theories, ‘conjunction’ here is naturally read as propositional conjunction, so that we should think of ew-accounts as something like very complex propositions and ‘conjunctions’ of them as literal conjunctions of those propositions. But this requires, on pain of experience (and thus the appearances it ontologically specifies) being inconsistent (if p and q are mutually inconsistent then p & q is inconsistent), that experience be ‘conservative,’ as Jauernig explicitly notes: ‘we must also assume that the evolution of our ew-accounts is conservative and cumulative so that there are no revisions or rejections of earlier ew-accounts’ (p. 95). This means that if one ew-account represents an object as having a determinate empirical property (e.g. a determinate mass) no later ew-account can represent it as lacking that property or having a distinct such property (e.g. a different mass). Likewise, if one ew-account represent a as existing, no later ew-account can represent a as non-existent (e.g. as the mistaken posit of a false theory). However, this account faces several serious difficulties. First, by grounding the actual empirical properties of objects in actual ew-accounts, Jauernig's account makes the actual empirical properties of objects dependent on the actual course of experience and, thus, effectively, on the actual future of science (given her ‘science-boosted’ conception of ew-accounts.) Assuming that climate change at some point wipes out all of humanity (and on the admittedly un-Kantian assumption that there is no rational life on other planets), Jauernig's view entails that the empirical world is no more determinate than will be represented in the most complete ew-account actually achieved before the climate apocalypse.14 Likewise, even if humanity survives, but, contrary to our rational vocation, simply stops doing science altogether (perhaps we all get addicted to video games), Jauernig's account entails that the empirical world is only as determinate as it is represented in the last ew-account before the great un-enlightenment. Not only is this intuitively absurd, it reduces Kant's view to the rejected Final Experience view from earlier (because our most complete scientific theory before the climate apocalypse will literally be final experience.) More nebulously, it also conflicts with the spirit of Kant's ‘scientific realism,’ i.e. his belief that the natural world exists and is determinate regardless of what we actually discover about it.15 This is one reason to explore counterfactual views like those briefly mentioned earlier.16 But perhaps Kant's idealism simply has these wildly counter-intuitive consequences, or he is simply not a scientific realist in this sense, or his idealism requires the immortality of the human species and of science, so I will consider other reasons to be dissatisfied with Jauernig's account. Second, Jauernig's definition of ew-accounts appeals to ‘known’ empirical laws.17 This uncharacteristic slip into epistemology conceals the fact that those laws are among the very things that need to be grounded as well. The transcendental principles of nature (e.g. the conservation of the quantity of substance) are completely grounded in the nature of our cognitive faculties; this is why they are a priori and pure. The metaphysical principles of matter are partly grounded in these transcendental principles, but they also partly depend on the empirical concept <matter>.18 Particular natural laws (e.g. chemical laws, or laws of matter that are not metaphysical principles) are also partly, but not wholly, grounded in metaphysical and transcendental principles. This is why they cannot be known fully a priori. Because Jauernig has built the natural laws into the grounding base for the empirical properties of objects, her account is not in a position to provide a ground for those very laws themselves. At various times she talks of them as ‘properly grounded’ in transcendental and metaphysical principles (p. 95; quoted below), but she never explicitly accounts for what this proper grounding amounts to. It cannot mean merely that they conform to transcendental and metaphysical principles, for the transcendental and metaphysical principles underdetermine the particular natural laws and thus do not constitute a complete ground for them. Thirdly, what entitles Jauernig (or Kant for that matter), to the assumption that experience must be conservative in this sense? Why can't it be that one ew-account represents an object as having a determinate empirical property, e.g. a determinate mass, but a later ew-account represents an object as having a slightly different determinate empirical property, e.g. a slightly greater or lower mass? Why can't the sequence of ever more complete and determinate ew-accounts successively revise one another, but represent a determinate empirical property (e.g. a determinate mass) because they converge to the assignment of that determinate empirical property to that object?19 Likewise, why can't one ew-account represent that some object exists (e.g. a planet in some distant solar system) but some later ew-account reject its existence (e.g. it was an artifact of bad telescope data), but the whole sequence represents the object as existing in virtue of there being a point after which all ew-accounts represent it as existing?20 One answer might be that we simply stipulate that the ew-account content within any putative experience, at any time, is the sub-set of the content that will not be rejected by any future experience. (If ‘stipulation’ sounds too arbitrary, think of this as ‘constructing’ the ew-accounts accordingly.) But this induces skepticism about the content of ew-accounts, for to claim that the content within an experience is part of its ew-account content is to claim that no future experience will reject or revise it. What empirical or a priori ground could I possibly have for claiming something so strong about what an ew-account represents about the empirical properties of an object (those that are not completely determined by metaphysical and transcendental principles)? We merely have to assume that (1) no presumed law that is properly grounded in the transcendental principles of the Critique is ever up for revision, and (2) no ew-account includes any descriptions of appearances or statements of presumed laws, be they scientific or commonsensical, that conflict with a transcendentally properly grounded law that has been, will be, or could be, specified by science, keeping fixed the transcendental level of reality. Of course, there probably will be many such descriptions and statements on the record at any given time; they just will not be part of the ew-account at that time. (p. 95) If ‘many such descriptions and statements [are] on the record at any given time’ but are not part of the ew-account at that time, then how can I know which of these descriptions and statements part of the ew-account at that time? One thing I can do is could do is wait and see what gets preserved in the ew-accounts of the future – but how do I know which descriptions and statement on the record in the future are part of the ew-account at that future time? I would have to know whether their proposed laws ‘conflict with a transcendentally properly grounded law that has been, will be, or could be, specified by science, keeping fixed the transcendental level of reality.’ But, aside from ruling out proposed laws that conflict with the transcendental and metaphysical principles of experience, this of no use. For I do not know whether the putative law ‘has been’ specified by science, because I do not know which parts of past science are contained in past ew-accounts. I can't look to the future because that just pushes the problem back a step (recall the dialectic: we are looking to the future to determine the ew-account of the present). And I can't know which particular laws ‘could’ be properly grounded ‘keeping fixed the transcendental level of reality’ because I do not have positive, substantive knowledge of that transcendental level of reality.21 I have put this point epistemically but it can also be put metaphysically: in virtue of what is some content in the descriptions and statement on the record at some time part of the ew-account at that time, rather than not? It cannot be grounded in the content of previous ew-accounts, on pain of circularity (ew-account content in the past is what gets retained in the future, i.e. the time in question). And it cannot be grounded in the future, on pain of regress. If it is grounded in what ‘could be specified by science, keeping fixed the transcendental level of reality’ then not only is it beyond our epistemic reach, but it also conflicts with the strong mind-dependence of appearances and their ‘ontological ingredients’ (see above).22 Fourthly, a mere conjunction of ew-accounts, even a conjunction of converging ew-accounts, does not represent determinate empirical properties of objects. Consider a sequence of ew-accounts each of which assigns the mass of the moon a determinate quantity, such that the sequence of these assignments converges to m. Now take the conjunction of all these assignments (i.e. the propositions that the moon has the assigned mass.) That conjunction is compatible both with the proposition that the moon has mass m and the proposition that the moon does not have mass m. A merely conjunctive account of how the representational content of experience depends upon the representational content of ew-accounts is not sufficient.23 In conclusion, while I agree with the general thrust of Jauernig's idealist reading of Kant, I am less convinced by her account of how experience ontologically specifies appearances.